On 18.10.2016 09:35, Karol Herbst wrote:> how sure are you, that this is needed for _every_ nvac? >Thank you for asking. If you consider, as relevant, referring to the original commit: "drm/nouveau/disp/g94: implement workaround for dvi issue on fx380" https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2a4bd8a <quote> Fixes the second DVI output on Quadro FX380. Thanks to NVIDIA for providing the details on the full workaround. [...] + switch (device->chipset) { + case 0x94: + case 0x96: + case 0x98: + case 0xaa: + case 0xac: + return true; [...] </quote> and to Quadro FX380 as defined: 1. https://nouveau.freedesktop.org/wiki/CodeNames/#NV50 NV96 (G96) ... 2. https://en.wikipedia.org/wiki/Nvidia_Quadro G96 ... GeForce 9400 based 3. https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#Quadro_FX_.28x800.29_series G96 ... The right question would be, for you Karol, Ben and perhaps the ones from the NVIDIA - those to which Ben refers, whether device->chipset: + case 0x94: + case 0x98: + case 0xaa: + case 0xac: are redundant, in the first place? Moreover, even if case 0x96 applies only, how sure are -you-, that this is needed for _every_ nv96? And given that I am here only the user, who is only caring for my hardware, I can only appreciate your sense of humor. ;)
well, I just don't want that this fix breaks the same thing for other users, that's why I am asking. 2016-10-18 13:56 GMT+02:00 poma <pomidorabelisima at gmail.com>:> On 18.10.2016 09:35, Karol Herbst wrote: >> how sure are you, that this is needed for _every_ nvac? >> > > Thank you for asking. > > If you consider, as relevant, > referring to the original commit: > "drm/nouveau/disp/g94: implement workaround for dvi issue on fx380" > https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=2a4bd8a > <quote> > Fixes the second DVI output on Quadro FX380. > Thanks to NVIDIA for providing the details on the full workaround. > > [...] > + switch (device->chipset) { > + case 0x94: > + case 0x96: > + case 0x98: > + case 0xaa: > + case 0xac: > + return true; > [...] > </quote> > > and to Quadro FX380 as defined: > > 1. https://nouveau.freedesktop.org/wiki/CodeNames/#NV50 > NV96 (G96) ... > > 2. https://en.wikipedia.org/wiki/Nvidia_Quadro > G96 ... GeForce 9400 based > > 3. https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#Quadro_FX_.28x800.29_series > G96 ... > > > The right question would be, > for you Karol, Ben and perhaps the ones from the NVIDIA - those to which Ben refers, > > whether device->chipset: > + case 0x94: > + case 0x98: > + case 0xaa: > + case 0xac: > are redundant, in the first place? > > Moreover, even if case 0x96 applies only, > how sure are -you-, that this is needed for _every_ nv96? > > And given that I am here only the user, who is only caring for my hardware, > I can only appreciate your sense of humor. ;) > > >
On 18.10.2016 16:02, Karol Herbst wrote:> well, I just don't want that this fix breaks the same thing for other > users, that's why I am asking. >Affected device ID: https://github.com/skeggsb/nouveau/blob/master/drm/nouveau/nvkm/engine/device/pci.c#L1229 can it be excluded from device->chipset case 0xac ? Care to create a patch? I'll test it.